Unverified50% confidenceFactExact time
LLM inference latency is typically broken down into Time To First Token (TTFT) and Time Per Output Token (TPOT).
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedTime Per Output Token (TPOT) measures the speed at which a model generates tokens one by one during the autoregressive decoding phase.77% similarVerifiedTTFT（首个token生成时间）是衡量LLM用户体验的核心指标，决定用户从提交请求到看到第一个字符的等待时长71% similarUnverifiedTIMESTAMP stores data internally in UTC, converting the session timezone to UTC on write and restoring it to the session timezone on read; DATETIME stores the input literal value as-is without any timezone conversion.67% similarUnverifiedTime to First Byte (TTFB) reflects the latency from when a request is sent to when the model returns the first token, encompassing network transmission, request queue waiting time, and model warm-up time.66% similarUnverifiedRepresentative tools for temporal expression normalization include SUTime and HeidelTime, but LLMs handle temporal expressions not by relying on these rule engines but by inferring from statistical patterns in training data.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/912164API
curl https://kongchang.com/api/v1/knowledge/claims/912164MCP
get_claim(id=912164)